A5033842671- Advanced Vision and Imaging
- Computer Graphics and Visualization Techniques
- Generative Adversarial Networks and Image Synthesis
- Visual Attention and Saliency Detection
- Image Enhancement Techniques
- Advanced Image and Video Retrieval Techniques
- Video Analysis and Summarization
- Image Retrieval and Classification Techniques
- 3D Shape Modeling and Analysis
- Human Motion and Animation
- 3D Surveying and Cultural Heritage
- Cinema and Media Studies
- Digital Media and Visual Art
- Cultural and Historical Studies
- Handwritten Text Recognition Techniques
- Image Processing and 3D Reconstruction
Vietnam National University Ho Chi Minh City
2024
Ho Chi Minh City International University
2024
National Cheng Kung University
2022-2024
Deploying style transfer methods on resource-constrained devices is challenging, which limits their real-world applicability. To tackle this issue, we propose using pruning techniques to accelerate various visual methods. We argue that typical may not be well-suited for and present an iterative correlation-based channel (ICCP) strategy encoder-transform-decoder-based image/video models. The regularization preserves the feature distributions content references, prevents layer collapse when...
Abstract The authors propose a deep neural network–based algorithm to automatically generate portrait map art (PMA), modern form created by British artist Ed Fairburn. formulate the generation of PMA as an adaptive dual-to-single image translation problem. authors’ proposed model analyzes appearance one and using two encoder networks utilizes their hidden encodings representations new decoder network. An style harmonization module is fuse encodings. Optimized cycle-consistency constraint,...
Changing the style of an image/video while preserving its content is a crucial criterion to access new neural transfer algorithm. However, it very challenging map art certain video in which “content” comprises background and animation objects. In this article, we present novel comprehensive system that solves problems transferring such video. Our takes as input arbitrary video, image, off-the-shelf image. It then generates artistic without damaging functionality consistency details. To solve...
Wire sculptures are objects sculpted by the use of wires. In this article, we propose practical methods to create 3D virtual wire sculptural art from a given model. contrast, most previous results reconstructed input 2D images. Artists usually tend design their with single if possible. If not possible, they try it least number To follow general trend, our proposed method generates minimum continuous lines. achieve goal, first adopt greedy approach extract important edges These extracted...
The neural mechanisms underlying hypnotic states and responses to suggestions remain largely unknown and, date, have been studied only with indirect methods. Here, the effects of hypnosis alter pain perception were ...
If the video has long been mentioned as a widespread visualization form, animation sequence in is storytelling for people. Producing an requires intensive human labor from skilled professional artists to obtain plausible both content and motion direction, incredibly animations with complex content, multiple moving objects, dense movement. This paper presents interactive framework generate new sequences according users' preference on starting frame. The critical contrast of our approach...
Image collage is a very useful tool for visualizing an image collection. Most of the existing methods and commercial applications generating collages are designed on simple shapes, such as rectangular circular layouts. This greatly limits use in some artistic creative settings. Although there that can generate irregularly-shaped collages, they often suffer from severe overlapping excessive blank space. prevents being effective information communication tools. In this paper, we present shape...
Video holds significance in computer graphics applications. Because of the heterogeneous digital devices, retargeting videos becomes an essential function to enhance user viewing experience such In research video retargeting, preserving relevant visual content videos, avoiding flicking, and processing time are vital challenges. Extending image techniques domain is challenging due high running time. Prior work mainly utilizes time-consuming preprocessing analyze frames. Plus, being tolerant...
From a single still image, looping video could be generated by imparting subtle motion to objects in the image. The results are hybrid of photography and video. They contain gentle some objects, while rest image remains still. Existing techniques successful animating such images. However, there drawbacks that need investigated, as too-large computation time necessary retrieve matched videos or challenges controlling desired not only terms region but also consistency regions. In this work, we...
No abstract available.
We propose a novel deep learning framework for animation video resequencing. Our system produces new sequences by minimizing perceptual distance of images from an existing clip. To measure distance, we utilize the activations convolutional neural networks and learn training these features on small network with data comprised human judgments. show that this metric graph-based manifold techniques, our can produce smooth visually appealing results variety styles. In contrast to previous work...